redundancy questionj

[A.L.M.Buxey at lboro.ac.uk]

Hi,

> Hello all, I'm new to FreeRADIUS and I'm planning to use Ver 2.2.0 on Ubuntu 12.04 server.
> Have 1 central office and 6 remote locations.  
> 1.  Is it standard to put one server in each of the remote locations which each remote back to the central office server?  Fail-over is a term used in the wiki, is that the same as redundancy?

there are many deployment models...what do YOU need?  If the central office is down, can the
remote locations operate autonomously? can authentication work if the remote office
is down?  if not, then what point having a local FR instance?  what is your bandwidth on the
WAN - if a little tight then a local caching FR server might help with reduction of some
traffic

> 2.  How stable is FreeRadius...should I consider a primary and secondary server in each remote location?  Or is that over kill?

'pretty stable' - it all depends on what you do with it. some people launch and leave...and
then maybe pick up the pieces (usually on this list) several years later when it has gone
wonky...and found the install undocumented etc..'it just worked for years' - others alter
their config randomly...not checking on a test/dev server first and break.  however, in the main,
once its up and running and working then it just does. you need to ensure you have correct
startup/recovery and monitoring scripts...tools like 'monit' can monitor your daemons and kick
them a few times if they have gone AWOL....disk space is usually the one that kills in a nasty
way.

alan